Lymphovascular and Perineural Invasion As Selection Criteria for Adjuvant Therapy in Intrahepatic Cholangiocarcinoma: a Multi-Institution Analysis

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector DOI:10.1111/j.1477-2574.2012.00489.x HPB

ORIGINAL ARTICLE Lymphovascular and perineural invasion as selection criteria for adjuvant therapy in intrahepatic cholangiocarcinoma: a multi-institution analysis

Sarah B. Fisher1, Sameer H. Patel1, David A. Kooby1, Sharon Weber2, Mark Bloomston3, Clifford Cho2, Ioannis Hatzaras3, Carl Schmidt3, Emily Winslow2, Charles A. Staley III1 & Shishir K. Maithel1

1Division of Surgical Oncology, Department of Surgery, Winship Cancer Institute, Emory University, Atlanta, GA, USA, 2Division of Surgical Oncology, Department of Surgery, University of Wisconsin, Madison, WI, USA, and 3Division of Surgical Oncology, Department of Surgery, Ohio State University, Columbus, OH, USA

Abstracthpb_489 514..522 Objectives: Criteria for the selection of patients for adjuvant chemotherapy in intrahepatic cholangiocarcinoma (IHCC) are lacking. Some authors advocate treating patients with lymph node (LN) involvement; however, nodal assessment is often inadequate or not performed. This study aimed to identify surrogate criteria based on characteristics of the primary tumour. Methods: A total of 58 patients who underwent resection for IHCC between January 2000 and January 2010 at any of three institutions were identified. Primary outcome was overall survival (OS). Results: Median OS was 23.0 months. Median tumour size was 6.5 cm and the median number of lesions was one. Overall, 16% of patients had positive margins, 38% had perineural invasion (PNI), 40% had lymphovascular invasion (LVI) and 22% had LN involvement. A median of two LNs were removed and a median of zero were positive. Lymph nodes were not sampled in 34% of patients. Lymphovascular and perineural invasion were associated with reduced OS [9.6 months vs. 32.7 months (P = 0.020) and 10.7 months vs. 32.7 months (P = 0.008), respectively]. Lymph node involvement indicated a trend towards reduced OS (10.7 months vs. 30.0 months; P = 0.063). The presence of either LVI or PNI in node-negative patients was associated with a reduction in OS similar to that in node-positive patients (12.1 months vs. 10.7 months; P = 0.541). After accounting for adverse tumour factors, only LVI and PNI remained associated with decreased OS on multivariate analysis (hazard ratio 4.07, 95% confidence interval 1.60–10.40; P = 0.003). Conclusions: Lymphovascular and perineural invasion are separately associated with a reduction in OS similar to that in patients with LN-positive disease. As nodal dissection is often not performed and the number of nodes retrieved is frequently inadequate, these tumour-specific factors should be considered as criteria for selection for adjuvant chemotherapy.

Keywords intrahepatic cholangiocarcinoma, lymphovascular invasion, perineural invasion, lymphadenectomy

Received 20 March 2012; accepted 17 April 2012

Correspondence Shishir K. Maithel, Division of Surgical Oncology, Department of Surgery, Winship Cancer Institute, Emory University School of Medicine, 1365C Clifton Road NE, Building C, 2nd Floor, Atlanta, GA 30322, USA. Tel: + 1 404 778 5777. Fax: + 1 404 778 4255. E-mail: [email protected]

This paper was presented at the American Hepato-Pancreato-Biliary Asso- ciation Annual Meeting, 7–12 March 2012, Miami Beach, Florida, and will be presented at the 10th World Congress of the International Hepato- Pancreato-Biliary Association, 1–5 July 2012, Paris.

Introduction institution series of resected hilar and intrahepatic CC.19 The current multi-institution study focuses specifically on prognostic Cholangiocarcinoma (CC) is the most common primary hepatic factors in resected IHCC. malignancy after hepatocellular carcinoma. It affects 5000–8000 individuals per year in the USA1 and globally accounts for 3% of all gastrointestinal malignancies.2 Cholangiocarcinoma is an Materials and methods aggressive cancer in which longterm survival is poor as a result Hepatobiliary resection databases at three institutions (Emory of the late presentation of disease and the limited therapies University, University of Wisconsin and Ohio State University) available. Indeed, the overall mortality rate in CC approaches its were reviewed for all patients with a diagnosis of IHCC incidence.1 who underwent resection between January 2000 and January Cholangiocarcinoma is divided into three general categories 2010. Permission from the institutional review board of according to whether the anatomic location of origin of disease is each institution was obtained prior to data review. Health intrahepatic, hilar or distal. Although hilar CC remains the most Insurance Portability and Accountability Act compliance was common type, the incidence of intrahepatic cholangiocarcinoma ensured. (IHCC) is rising3–5 and IHCC currently accounts for 20% of all A total of 58 patients who underwent resection for IHCC were CC.5 Many believe that the anatomic location of CC corresponds identified. Thirty-five of these patients had been included in a to specific and distinct tumour biology, as evidenced by separate prior analysis that assessed the value of LVI in a single-institution staging systems for intrahepatic, hilar and distal tumours; there- series of intrahepatic and hilar CC.19 Data for these 35 patients fore, each type of CC should be considered separately. This study are now reanalysed with updated follow-up, along with data for a focuses specifically on IHCC. different cohort of patients who underwent treatment for IHCC Resection is the mainstay of treatment for IHCC and offers at two other institutions. Overall survival was ascertained the only opportunity for cure. Resectability rates in IHCC range through the clinical follow-up documented in each patient’s from 46% to 75%.6 Improved surgical technique and refinements medical record and the Social Security Death Index. Pathology in patient selection and perioperative care have contributed reports were reviewed for important tumour factors that are to increased survival in resectable IHCC.7 Unfortunately, even known to have prognostic value for patient survival. These after complete resection, 5-year survival in IHCC remains poor include tumour size, number and grade, margin status, LN (5–43%), indicating that resection alone is not sufficient for most involvement, and the presence of LVI or perineural invasion patients.1,8 (PNI).15,20,21 Patients who did not undergo LN dissection were Current chemotherapy regimens are only marginally effective regarded as node-negative as this is consistent with clinical prac- and gemcitabine monotherapy has not been shown to be benefi- tice patterns. Perioperative mortality was defined as death within cial in the adjuvant setting.9 No guidelines for adjuvant therapy 90 days of operation. for IHCC exist.10 Recently, the Advanced Biliary Cancer (ABC) Trial demonstrated improved survival in patients with advanced- stage disease who received doublet chemotherapy, consisting of Statistical analysis gemcitabine and cisplatin, compared with gemcitabine mono- Data were analysed using spss Version 17.0 for Windows (SPSS, therapy (11.7 months vs. 8.1 months; P < 0.001).11 It is possible Inc., Chicago, IL, USA). Kaplan–Meier log-rank survival analysis that the survival advantage of this more aggressive doublet was used to determine the association of each pathologic factor regimen may translate to the adjuvant setting, but this must be with patient survival. Univariate Cox regression analysis was per- evaluated in a prospective clinical trial. The benefit of adjuvant formed for adverse tumour factors to determine their association therapy in resected IHCC will undoubtedly depend on appropri- with OS. Factors significant at a level of P < 0.2 were included in ate patient selection. a multivariate Cox regression model. The presence of lymph node (LN) metastases is commonly utilized to select patients for adjuvant therapy after complete Results resection of many gastrointestinal malignancies, including IHCC.12–15 However, portal lymphadenectomy is not routinely A total of 58 patients underwent resection. The median age of the performed in IHCC and its use remains controversial.6,8,16–18 Fur- study sample was 66 years (range: 29–89 years); 38 patients (66%) thermore, when it is performed, its nodal yield is often inadequate were female. Pathologic and perioperative variables are summa- for accurate staging. Given the frequent lack of information rized in Table 1. Thirteen patients had pathologically proven LN regarding nodal status in IHCC, the present study sought to iden- disease. Rates of LN positivity ranged from 22% in all patients in tify a surrogate marker for adverse tumour biology based on the this study to 34% in those who underwent LN procurement with characteristics of the primary tumour. The authors have previ- pathologic assessment (n = 38). Because of the referral patterns of ously shown lymphovascular invasion (LVI) to be an independent the three institutions, data on adjuvant therapy are limited. Of the prognostic factor for shortened overall survival (OS) in a single- 48 patients whose adjuvant treatment status is known, 13 patients

Table 1 Clinicopathologic characteristics of patients with intrahe- 81.4 months). The results of univariate and multivariate Cox patic cholangiocarcinoma (n = 58) regression analyses for OS are shown in Table 2. Variable Patients, n (%) Median (range) Lymph node involvement resulted in a strong trend towards Operative characteristics reduced OS (10.7 months vs. 30.0 months; P = 0.063) (Fig. 1a). A Estimated blood loss, ml 500 (50–4000) subset analysis of patients who did not undergo lymphadenec- Type of resection tomy compared with those who were pathologically identified as Left hepatectomy 14 (24%) LN-negative did not demonstrate any difference in survival Left lateral sectorectomy 6 (10%) (54.4 months vs. 21.1 months; P = 0.234) (Fig. 1b). Lymphovas- cular invasion was associated with reduced OS (9.6 months vs. Left trisegmentectomy 1 (2%) 32.7 months; P = 0.020) (Fig. 2a). Perineural invasion was also Right hepatectomy 8 (14%) associated with reduced OS (10.7 months vs. 32.7 months; P = Right trisegmentectomy 10 (17%) 0.008) (Fig. 2b). A representative image demonstrating LVI and Extended right 13 (23%) PNI is shown in Fig. 3. hepatectomy Lymphovascular and perineural invasion were statistically Other 6 (10%) more likely to occur within the same tumour specimen [LVI+/ Pathologic characteristics PNI+: 24.1%; LVI+/PNI-: 15.5%; LVI-/PNI+: 13.8%; LVI-/PNI-: Positive margin 9 (16%) 46.6% (P = 0.008)], but the presence of both LVI and PNI did not Tumour size, cm 6.5 (1.3–21) portend worse survival compared with the presence of either Tumour size Ն6.5 cm 32 (55%) alone (Fig. 4). There was no association between LN positivity and Number of lesions 1 (1–7) the presence of LVI [LN+/LVI+: 12.1%; LN-/LVI+: 27.6% (P = Multiple tumours 12 (21%) 2 (2–7) 0.387)] or between LN positivity and the presence of PNI [LN+/ No lymphadenectomy 20 (34%) PNI+: 13.8%; LN-/PNI+: 24.1% (P = 0.095)]. In a subset analysis performed of node-negative patients only (n = 45), LVI demonstrated a Lymph node-positive disease 13 (22%) strong trend towards reduced survival (12.1 months vs. Number of nodes retrieved 2 (1–10) 45.8 months; P = 0.124) (Fig. 5a). In the same subset of Number of positive nodes 0 (0–5) LN-negative patients, PNI was significantly associated with Differentiation reduced OS (12.1 months vs. 32.7 months; P = 0.052) (Fig. 5b). In Good 3 (5%) node-negative patients, the presence of either LVI or PNI was Moderate 32 (55%) associated with reduced OS (12.1 months) similar to that in patients with node-positive disease (10.7 months) (P = 0.541) Poor 19 (33%) (Fig. 6). After accounting for adverse tumour factors including Unknown 4 (7%) tumour size and number, margin status, grade, and LN status, Lymphovascular invasion 23 (40%) only the presence of LVI or PNI persisted as negative prognostic Perineural invasion 22 (38%) factors for reduced OS in a multivariate Cox regression analysis Postoperative course (hazard ratio 4.07, 95% confidence interval 1.60–10.40; P = 0.003) Length of hospital stay, days 8 (1–44) (Table 2). Complications 30 (52%) Infectious 15 (26%) Discussion Bleeding 3 (5%) Bile leak 3 (5%) Cholangiocarcinoma is a deadly disease with an overall mortality 22 Reoperation 3 (5%) rate that is similar to its incidence. Resection offers the only Perioperative 30-day mortality 6 (10%) chance for cure, yet 5-year survival in this population remains < 40% in most studies.23–25 Despite this dismal prognosis, evidence Perioperative 90-day mortality 8 (14%) supporting the use of adjuvant chemotherapy after complete resection of IHCC is lacking.10,15,26 Recently, the ABC-02 Trial demonstrated an improvement in OS in the advanced disease received chemotherapy and six received radiation therapy. Precise setting with a doublet chemotherapy regimen of gemcitabine and information on the chemotherapy regimen was not available. cisplatin.11 It remains to be determined whether or not this advantage will transfer to the adjuvant setting. As with other malignan- Survival analysis cies, appropriate selection of patients for adjuvant therapy will be The median follow-up in survivors was 22.0 months (range: 0.4– key to maximizing therapeutic benefits while minimizing toxicity. 81.4 months). At the time of last follow-up, 35 patients (60%) had The presence of adverse pathologic factors, such as LN died. Median OS in all patients was 23.0 months (range: 0.3– metastases, is commonly utilized to select patients for adjuvant

Table 2 Factors associated with overall survival (90-day mortality excluded, n = 50) Variable Univariate analysis Multivariate analysis

HR 95% CI P-value HR 95% CI P-value Positive margin 0.59 0.17–2.00 0.394 – – – Tumour size Ն6.5 cm 1.11 0.49–2.54 0.805 – – – Multiple tumours 1.73 0.77–3.90 0.188 0.91 0.35–2.35 0.839 Poor grade 2.23 0.92–5.42 0.077 2.52 0.94–6.77 0.066 Positive lymph node involvement 2.45 1.05–5.74 0.039 1.90 0.72–5.03 0.197 LVI or PNI 3.39 1.52–7.56 0.003 4.07 1.60–10.40 0.003

HR, hazard ratio; 95% CI, 95% confidence interval; LVI, lymphovascular invasion; PNI, perineural invasion.

1.0 1.0 Node-negative (n = 45) Nx (n = 20) Node-positive (n = 13) N0 (n = 25)

0.8 0.8 P = 0.063 P = 0.234

0.6 0.6 54.4 months

30.0 months 0.4 0.4 Proportion surviving Proportion 0.2 0.2 21.1 months 10.7 months

0.0 0.0

0.0 10.0 20.0 30.0 40.0 50.0 60.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 25 15 9 3 12 10 6 2

51 11 13 5 3 1 (a) (b) Time, months Time, months Figure 1 Nodal status and overall survival in (a) node-negative (Nx + N0) vs. node-positive patients, and (b) clinically node-negative (Nx) vs. pathologically node-negative (N0) patients therapy after complete resection of many gastrointestinal undergo lymphadenectomy are considered to be LN-negative in malignancies.12–14 In IHCC, regional nodal involvement has been decisions on adjuvant therapy. The current study is also subject to shown to be a prognostic factor for reduced OS.6,17,18,27,28 The 7th this limitation, but it does represent an accurate reﬂection of edition of the American Joint Committee on Cancer (AJCC) clinical practice, particularly as lymphadenectomy is not a stan- staging system for IHCC takes into account LN involvement29 as it dard of care in IHCC. A subset analysis of patients who did not correlates well with OS.17,30 In the present study, LN positivity undergo lymphadenectomy compared with those who were ranged from 22% in all patients to 34% in the subset of patients pathologically LN-negative did not demonstrate any difference in who underwent LN procurement with pathologic assessment survival. In fact, patients in whom LN status was unknown (n = 38). This is consistent with a recent multi-institution series displayed a trend towards improved survival (Fig. 1b). Thus, it that reported LN involvement in the range of 16–30%.17 is unlikely that many of these patients harboured occult LN In Western countries, however, portal lymphadenectomy is metastases. not routinely performed for IHCC and its value remains Furthermore, when portal LNs are removed, the median controversial.6,8,16–18 In the clinical setting, patients who do not number of nodes retrieved is generally fewer than three.17–19 This

1.0 1.0 LVI-negative (n = 35) PNI-negative (n = 36) LVI-positive (n = 23) PNI-positive (n = 22)

0.8 0.8 P = 0.020 P = 0.008

0.6 0.6

32.7 months 32.7 months 0.4 0.4 Proportion surviving Proportion 0.2 0.2 9.6 months 10.7 months

0.0 0.0

0.0 10.0 20.0 30.0 40.0 50.0 60.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 23 12 8 3 22 13 9 3

74 21 83 11 (a) (b) Time, months Time, months Figure 2 Overall survival in patients with and without (a) lymphovascular invasion (LVI) and (b) perineural invasion (PNI)

Figure 3 Histopathology shows (a) lymphovascular invasion as demonstrated by tumour cells (black arrow) within a vascular channel and (b) perineural invasion with tumour cells (white arrow) within a neural bundle (*). (Haematoxylin and eosin stain; original magnification ¥400)

occurred in the present study, in which the median number of limitations of LN evaluation in IHCC, the aim of the present study nodes retrieved was only two. There are no speciﬁcations within was to assess characteristics of the primary tumour that might the AJCC guidelines as to the optimal number of nodes that must provide prognostic information similar to that indicated by LN be retrieved for accurate staging, but it is doubtful that such a involvement to select patients for adjuvant therapy. small number of nodes will enable a complete and accurate evalu- The present authors have previously demonstrated that the ation. Ito et al. recently reported that the retrieval of six LNs was presence of LVI is an independent prognostic factor for reduced necessary for accurate staging of extrahepatic CC.31 Given these OS based on a single-institution series of resected CC, which

LVI- and PNI-negative (n = 27) In a small single-institution study of only 22 patients with 1.0 40 LVI or PNI (n = 17) pathologically node-negative IHCC, Shirabe and colleagues LVI and PNI (n = 14) showed a high microvessel count (which is likely to represent a 0.8 surrogate for LVI) was an independent prognostic factor for poor OS. In the current study, patients who did not undergo portal lymphadenectomy were considered to have LN-negative disease as 0.6 this is a true representation of clinical practice patterns. Given the 54.4 months rarity of IHCC and the heterogeneity in the performance of LN dissections, a study to assess only histologically assessed LNs is 0.4 difﬁcult outside the context of a prospective trial. A subset analysis 11.9 months of clinically and pathologically node-negative patients (n = 45) in

Proportion surviving Proportion the current study demonstrated a strong trend towards reduced 0.2 P = 0.002 9.2 months P = 0.842 survival in patients with LVI (12.1 months vs. 45.8 months; P = 0.124) (Fig. 5a). However, this subset analysis is somewhat 0.0 limited by the number of patients. Like LVI, PNI has been implicated as a poor prognostic 0.0 10.0 20.0 30.0 40.0 50.0 60.0 factor in many cancer types, including squamous cell 19 12 8 3 carcinoma of the head and neck, cancer of the prostate, and 7111 colorectal and pancreatic cancers.41–45 In a large series of hilar 4311 and intrahepatic CC, Endo and colleagues showed that the Time, months presence of PNI was significantly associated with reduced Figure 4 Overall survival in patients with and without lymphovascu- OS.46 Several smaller single-institution series that included lar invasion (LVI) and perineural invasion (PNI) shows that the effects only patients with IHCC have demonstrated a similar associa- of LVI and PNI are not additive tion between PNI and shortened survival, but failed to show the significance of PNI when accounting for other adverse pathologic factors.8,47 In the authors’ recent single-institution included both hilar and intrahepatic disease sites.19 One of the study, which included patients with both hilar and intrahepatic major limitations of that study was its inclusion of both hilar and CC, PNI marked a trend towards reduced OS, but this did not intrahepatic disease sites, which many regard as separate entities reach statistical significance.19 In the current multi-institution with distinct tumour biology. The present study is a subsequent study on IHCC, the presence of PNI was found to be signifi- investigation that focused only on IHCC. It included 35 patients cantly associated with a worse prognosis, even after accounting from the original study, analysed using updated follow-up data, for other known adverse factors, such as large tumour size, mul- and a different cohort of patients treated at two other institutions. tiple tumours, positive resection margin, poor differentiation In the current study, which is specific to IHCC, the presence of LVI and LN involvement. Although multiple tumours might be and/or PNI confers a negative prognostic effect on survival similar expected to represent poor tumour biology and thus indicate to that implied by LN involvement. The authors propose that poor survival, most patients included in this series had solitary these characteristics of the primary tumour may be a surrogate for tumours. Of the 12 patients with multiple lesions, more than aggressive tumour behaviour and may serve as criteria for select- half (n = 7) had only two tumours. This fact, along with the ing for adjuvant therapy. Additionally, the presence of LVI and/or inherent difficulty of capturing the process of careful patient PNI may be used to stratify patient populations in future studies selection for operation in a retrospective study, is likely to assessing the efficacy of adjuvant regimens. explain the lack of prognostic significance of multiple tumours The concept of LVI as an adverse prognostic marker is not in this series. unique to CC. Tumoral lymphovascular involvement signifi- In the current study, lymphovascular and perineural invasion cantly correlates with rates of LN metastases in breast cancer,32,33 demonstrated a parallel relationship in that, in a given patient, endometrial cancer34,35 and colon cancer36 and is regarded as a positivity or negativity for one was likely to be reflected by, poor prognostic factor for survival.37,38 In the current series, the respectively, positivity or negativity for the other. Studies of lym- presence of LVI did not correlate with the presence of nodal phatic and blood vessel invasion and PNI in gastric cancer disease. Although this may indicate inadequate LN sampling in patients have shown similar findings.48–50 The presence of both IHCC, as previously discussed, it does reflect clinical reality and factors, however, did not portend a prognosis worse than that supports the suggestion that LVI is an independent negative implied by the presence of either alone (9.2 months vs. prognostic factor. Studies in node-negative breast cancer support 11.9 months; P = 0.842) (Fig. 4). Thus, LVI and PNI both sepa- the concept of LVI as an independent poor prognostic factor for rately and in combination confer equal but independent negative OS.39 prognostic effects on survival.

1.0 1.0 LVI-negative (n = 29) PNI-negative (n = 31) LVI-positive (n = 16) PNI-positive (n = 14) P = 0.124 P = 0.052 0.8 0.8

0.6 0.6

45.8 months 32.7 months

0.4 0.4 Proportion surviving Proportion 0.2 12.1 months 0.2 12.1 months

0.0 0.0

0.0 10.0 20.0 30.0 40.0 50.0 60.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 19 11 7 2 19 12 8 2

64 21 63 11 (a) (b) Time (truncated to 60 months) Time (truncated to 60 months) Figure 5 Overall survival in patients without lymph node involvement (n = 45) with and without (a) lymphovascular invasion (LVI) and (b) perineural invasion (PNI)

Both LVI and PNI are routinely assessed in standard pathologic evaluations of resected IHCC. As characteristics of the primary 1.0 Node-, LVI- and PNI tumour, LVI and PNI are not subject to the variability of LN -negative (n = 25) retrieval. The presence of either LVI or PNI is associated with Node-negative, LVI- shortened OS equivalent to that in LN-positive disease. Given the or PNI-positive (n = 20) 0.8 prospect of improved and more efﬁcacious chemotherapy with Node-positive (n = 13) gemcitabine and cisplatin, the presence of LVI or PNI may represent a reproducible and reliable criterion for selecting patients for 0.6 adjuvant therapy after resection of IHCC. 45.8 months

0.4 Conflicts of interest None declared. 12.1 months Proportion surviving Proportion 0.2 P = 0.019 References 10.7 months P = 0.541 1. Shaib Y, El-Serag HB. (2004) The epidemiology of cholangiocarcinoma. Semin Liver Dis 24:115–125. 0.0 2. Khan SA, Toledano MB, Taylor-Robinson SD. (2008) Epidemiology, risk 0.0 10.0 20.0 30.0 40.0 50.0 60.0 factors, and pathogenesis of cholangiocarcinoma. HPB 10:77–82. 3. 17 11 7 2 Malhi H, Gores GJ. (2006) Cholangiocarcinoma: modern advances in understanding a deadly old disease. J Hepatol 45:856–867. 8421 5111 4. Endo I, House MG, Klimstra DS, Gönen M, D'Angelica M, Dematteo RP et al. (2008) Clinical significance of intraoperative bile duct margin Time, months assessment for hilar cholangiocarcinoma. Ann Surg Oncol 15:2104– Figure 6 Overall survival in patients with lymphovascular invasion or 2112. perineural invasion resembles that in patients with node-positive 5. McGlynn KA, Tarone RE, El-Serag HB. (2006) A comparison of trends in disease the incidence of hepatocellular carcinoma and intrahepatic cholangiocar-

cinoma in the United States. Cancer Epidemiol Biomarkers Prev 15:1198– hepatic resection: univariate and multivariate analysis. Hepatogastroen- 1203. terology 49:311–316. 6. Nakagawa T, Kamiyama T, Kurauchi N, Matsushita M, Nakanishi K, 21. Guglielmi A, Ruzzenente A, Campagnaro T, Pachera S, Valdegamberi A, Kamachi H et al. (2005) Number of lymph node metastases is a signifi- Nicoli P et al. (2009) Intrahepatic cholangiocarcinoma: prognostic factors cant prognostic factor in intrahepatic cholangiocarcinoma. World J Surg after surgical resection. World J Surg 33:1247–1254. 29:728–733. 22. Gatto M, Bragazzi MC, Semeraro R, Napoli C, Gentile R, Torrice A et al. 7. Nathan H, Pawlik TM, Wolfgang CL, Choti MA, Cameron JL, Schulick RD. (2010) Cholangiocarcinoma: update and future perspectives. Dig Liver (2007) Trends in survival after surgery for cholangiocarcinoma: a 30-year Dis 42:253–260. population-based SEER database analysis. J Gastrointest Surg 11:1488– 23. Jarnagin WR, Shoup M. (2004) Surgical management of cholangiocarci- 1496; discussion 1496–1497. noma. Semin Liver Dis 24:189–199. 8. Choi SB, Kim KS, Choi JY, Park SW, Choi JS, Lee WJ et al. (2009) The 24. Aljiffry M, Walsh MJ, Molinari M. (2009) Advances in diagnosis, treatment prognosis and survival outcome of intrahepatic cholangiocarcinoma fol- and palliation of cholangiocarcinoma: 1990–2009. World J Gastroenterol lowing surgical resection: association of lymph node metastasis and 15:4240–4262. lymph node dissection with survival. Ann Surg Oncol 16:3048–3056. 25. DeOliveira ML, Cunningham SC, Cameron JL, Kamangar F, Winter JM, 9. Todoroki T. (2000) Chemotherapy for bile duct carcinoma in the light of Lillemoe KD et al. (2007) Cholangiocarcinoma: thirty-one-year experience adjuvant chemotherapy to surgery. Hepatogastroenterology 47:644–649. with 564 patients at a single institution. Ann Surg 245:755–762. 10. Alvaro D, Cannizzaro R, Labianca R, Valvo F, Farinati F. (2010) Cholan- 26. Murakami Y, Uemura K, Sudo T, Hayashidani Y, Hashimoto Y, Nakamura giocarcinoma: a position paper by the Italian Society of Gastroenterology H et al. (2009) Gemcitabine-based adjuvant chemotherapy improves sur- (SIGE), the Italian Association of Hospital Gastroenterology (AIGO), the vival after aggressive surgery for hilar cholangiocarcinoma. J Gastrointest Italian Association of Medical Oncology (AIOM) and the Italian Associa- Surg 13:1470–1479. tion of Oncological Radiotherapy (AIRO). Dig Liver Dis 42:831–838. 27. Shimada K, Sano T, Nara S, Esaki M, Sakamoto Y, Kosuge T et al. (2009) 11. Valle J, Wasan H, Palmer DH, Cunningham D, Anthoney A, Maraveyas A Therapeutic value of lymph node dissection during hepatectomy in et al. (2010) Cisplatin plus gemcitabine versus gemcitabine for biliary patients with intrahepatic cholangiocellular carcinoma with negative tract cancer. N Engl J Med 362:1273–1281. lymph node involvement. Surgery 145:411–416. 12. Chang GJ, Rodriguez-Bigas MA, Skibber JM, Moyer VA. (2007) Lymph 28. Yamamoto M, Takasaki K, Yoshikawa T. (1999) Lymph node metastasis in node evaluation and survival after curative resection of colon cancer: intrahepatic cholangiocarcinoma. Jpn J Clin Oncol 29:147–150. systematic review. J Natl Cancer Inst Monogr 99:433–441. 29. Nathan H, Pawlik TM. (2010) Staging of intrahepatic cholangiocarcinoma. 13. Slidell MB, Chang DC, Cameron JL, Wolfgang C, Herman JM, Schulick Curr Opin Gastroenterol 26:269–273. RD et al. (2008) Impact of total lymph node count and lymph node 30. Farges O, Fuks D, Le Treut YP, Azoulay D, Laurent A, Bachellier P et al. ratio on staging and survival after pancreatectomy for pancreatic adeno- (2010) AJCC 7th edition of TNM staging accurately discriminates out- carcinoma: a large, population-based analysis. Ann Surg Oncol 15:165– comes of patients with resectable intrahepatic cholangiocarcinoma: by 174. the AFC-IHCC-2009 study group. Cancer 117:2170–2177. 14. Smith DD, Schwarz RR, Schwarz RE. (2005) Impact of total lymph 31. Ito K, Ito H, Allen PJ, Gonen M, Klimstra D, D'Angelica MI et al. (2010) node count on staging and survival after gastrectomy for gastric Adequate lymph node assessment for extrahepatic bile duct adenocar- cancer: data from a large US-population database. J Clin Oncol 23:7114– cinoma. Ann Surg 251:675–681. 7124. 32. Cornwell LB, McMasters KM, Chagpar AB. (2011) The impact of lym- 15. Murakami Y, Uemura K, Sudo T, Hashimoto Y, Nakashima A, Kondo N phovascular invasion on lymph node status in patients with breast et al. (2011) Prognostic factors after surgical resection for intrahepatic, cancer. Am Surg 77:874–877. hilar, and distal cholangiocarcinoma. Ann Surg Oncol 18:651–658. 33. Song YJ, Shin SH, Cho JS, Park MH, Yoon JH, Jegal YJ. (2011) The role 16. Ercolani G, Grazi GL, Ravaioli M, Grigioni WF, Cescon M, Gardini A et al. of lymphovascular invasion as a prognostic factor in patients with lymph (2004) The role of lymphadenectomy for liver tumours: further consider- node-positive operable invasive breast cancer. J Breast Cancer 14:198– ations on the appropriateness of treatment strategy. Ann Surg 239:202– 203. 209. 34. Guntupalli SR, Zighelboim I, Kizer NT, Zhang Q, Powell MA, Thaker PH 17. de Jong MC, Nathan H, Sotiropoulos GC, Paul A, Alexandrescu S, et al. (2012) Lymphovascular space invasion is an independent risk factor Marques H et al. (2011) Intrahepatic cholangiocarcinoma: an international for nodal disease and poor outcomes in endometrioid endometrial multi-institutional analysis of prognostic factors and lymph node assess- cancer. Gynecol Oncol 124:31–35. ment. J Clin Oncol 29:3140–3145. 35. Narayan K, Khaw P, Bernshaw D, Mileshkin L, Kondalsamy- 18. Clark CJ, Wood-Wentz CM, Reid-Lombardo KM, Kendrick ML, Huebner Chennakesavan S. (2012) Prognostic significance of lymphovascular M, Que FG. (2011) Lymphadenectomy in the staging and treatment of space invasion and nodal involvement in intermediate- and high-risk intrahepatic cholangiocarcinoma: a population-based study using the endometrial cancer patients treated with curative intent using surgery and National Cancer Institute SEER database. HPB13:612–620. adjuvant radiotherapy. Int J Gynecol Cancer 22:260–266. 19. Patel SH, Kooby DA, Staley CA, 3rd, Sarmiento JM, Maithel SK. (2011) 36. Royston D, Jackson DG. (2009) Mechanisms of lymphatic metastasis in The prognostic importance of lymphovascular invasion in cholangiocar- human colorectal adenocarcinoma. J Pathol 217:608–619. cinoma above the cystic duct: a new selection criterion for adjuvant 37. Gondo T, Nakashima J, Ozu C, Ohno Y, Horiguchi Y, Namiki K et al. therapy? HPB 13:605–611. (2011) Risk stratification of survival by lymphovascular invasion, patho- 20. Hanazaki K, Kajikawa S, Shimozawa N, Shimada K, Hiraguri M, Koide N logical stage, and surgical margin in patients with bladder cancer treated et al. (2002) Prognostic factors of intrahepatic cholangiocarcinoma after with radical cystectomy. Int J Clin Oncol. Epub 2011 Sept 7.

38. Brunocilla E, Pernetti R, Martorana G. (2011) The prognostic role of 45. Solares CA, Lee K, Parmar P, O'Rourke P, Panizza B. (2012) Epidemiol- lymphovascular invasion in urothelial-cell carcinoma of upper and lower ogy of clinical perineural invasion in cutaneous squamous cell carcinoma urinary tract. Anticancer Res 31:3503–3506. of the head and neck. Otolaryngol Head Neck Surg 146:746–751. 39. de Mascarel I, Bonichon F, Durand M, Mauriac L, MacGrogan G, Soubey- 46. Endo I, Gonen M, Yopp AC, Dalal KM, Zhou Q, Klimstra D et al. ran I et al. (1998) Obvious peritumoral emboli: an elusive prognostic (2008) Intrahepatic cholangiocarcinoma: rising frequency, improved factor reappraised. Multivariate analysis of 1320 node-negative breast survival, and determinants of outcome after resection. Ann Surg 248:84– cancers. Eur J Cancer 34:58–65. 96. 40. Shirabe K, Shimada M, Tsujita E, Aishima S, Maehara S, Tanaka S et al. 47. Chen TC, Jan YY, Yeh TS. (2012) K-ras mutation is strongly associated (2004) Prognostic factors in node-negative intrahepatic cholangiocarci- with perineural invasion and represents an independent prognostic factor noma with special reference to angiogenesis. Am J Surg 187:538–542. of intrahepatic cholangiocarcinoma after hepatectomy. Ann Surg Oncol. 41. Chatterjee D, Katz MH, Rashid A, Aishima S, Maehara S, Tanaka S et al. Epub 2012 Jan 31. (2012) Perineural and intraneural invasion in post-therapy pancreati- 48. Scartozzi M, Galizia E, Verdecchia L, Berardi R, Graziano F, Catalano V coduodenectomy specimens predicts poor prognosis in patients with et al. (2006) Lymphatic, blood vessel and perineural invasion identifies pancreatic ductal adenocarcinoma. Diagn Mol Pathol 36:409–417. early-stage high-risk radically resected gastric cancer patients. Br J 42. Katz B, Srougi M, Dall'oglio M, Nesrallah AJ, Sant'anna AC, Pontes J Cancer 95:445–449. et al. (2011) Perineural invasion detection in prostate biopsy is related to 49. Du C, Zhou Y, Cai H, Zhao G, Fu H, Shi YQ. (2012) Poor prognostic recurrence-free survival in patients submitted to radical prostatectomy. factors in patients with stage I gastric cancer according to the seventh Urol Oncol. Epub 2011 July 25. edition TNM classification: a comparative analysis of three subgroups. 43. Liebig C, Ayala G, Wilks JA, Berger DH, Albo D. (2009) Perineural invasion J Surg Oncol 105:323–328. in cancer: a review of the literature. Cancer 115:3379–3391. 50. Bilici A, Seker M, Ustaalioglu BB, Kefeli U, Yildirim E, Yavuzer D et al. 44. Peng J, Sheng W, Huang D, Venook AP, Xu Y, Guan Z et al. (2011) (2010) Prognostic significance of perineural invasion in patients with Perineural invasion in pT3N0 rectal cancer: the incidence and its prog- gastric cancer who underwent curative resection. Ann Surg Oncol nostic effect. Cancer 117:1415–1421. 17:2037–2044.